Pulse or digital communications – Bandwidth reduction or expansion – Television or motion video signal
Reexamination Certificate
2001-03-02
2003-11-18
Diep, Nhon (Department: 2613)
Pulse or digital communications
Bandwidth reduction or expansion
Television or motion video signal
C375S240250
Reexamination Certificate
active
06650707
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates in general to digital video signal processing and, more particularly, to improvements in video transcoding that reduce computation time.
2. Description of the Related Art
The present invention relates to the field of digital video signal processing. While definitions for selected terms in the field of digital video signal processing used herein are provided, a more complete set of definitions for this field of art is provided in International Standard ISO/IEC 13818-2 entitled “Information technology-Generic coding of moving pictures and associated audio information: Video,” dated May 15, 1996, and ITU-T Rec. H.263 entitled “Video coding for low bit-rate communication,” dated May 1997, which are incorporated in their entirety herein by reference. Also, the general requirements and operations of a video encoder are disclosed in Sun et al., “Statistical Computation of Discrete Cosine Transform in Video Encoders,” Journal of Visual Communications and Image Representation, Vol. 9, No. 2, pp. 163-170, June, 1998, which is incorporated in its entirety herein by reference.
Transcoding as used herein refers to an operation of decoding into the pixel domain an incoming encoded compressed video signal, and then re-encoding the decoded video signal into an encoded signal having a desired bit rate or format. Such transcoding enables, for example, converting an incoming video signal at a given bit rate into another data stream having a different bit rate. A more specific example of an application of transcoding is to convert the bit rate of an MPEG compressed video signal at, e.g., 9 Mbit/sec, to a lower bit rate so the signal can be relayed at a cable head with a limited capacity and lower bit rate.
FIG. 1
illustrates a cascaded pixel domain transcoder architecture
100
that includes a decoder section
102
and an encoder section
104
. Decoder section
102
includes a block
110
for performing an inverse quantization (IQ) on an incoming bit stream from a front-encoder (not shown) and a block
112
for performing an inverse discrete cosine transform (IDCT) on the output of block
110
. Block
112
outputs the decoded bit stream which is applied to a first input of a summer
114
. Motion compensation of the decoded bit stream is performed by a frame buffer (F)
116
and a motion compensation (MC) unit
118
coupled between a sum output and a second input of summer
114
and also coupled to receive motion vectors (MV).
Encoder section
104
includes a summer
120
having a first input coupled to receive the decoded and motion compensated bit stream from decoder section
102
. A block
122
performs a discrete cosine transform (DCT) on a sum output of summer
120
. The output of block
122
is applied to a block
124
which performs a quantization (Q) thereon. The output of block
124
is the encoded output of encoder section
104
and transcoder
100
. The encoded output is ultimately intended for receipt by an end-decoder (not shown).
Motion compensation of the encoded output of encoder section
104
is performed by a block
126
for performing an IQ, a block
128
for performing an IDCT, a summer
130
, a frame buffer (F)
132
and a motion compensation unit (MC)
134
, all coupled between the output of block
124
and a second input of summer
120
. Unit
134
is coupled to receive motion vectors (MV).
A direct implementation of transcoder
100
is to fully decode an incoming compressed bit-stream into the pixel-domain and then re-encode the decoded video into the desirable bit-rate. The cascaded pixel-domain transcoder architecture of transcoder
100
is flexible, since decoder
102
and encoder
104
can be totally independent of each other. For example, decoder
102
and encoder
104
can operate at different bit rates, picture resolutions, coding modes, and even according to different standards. The architecture can be implemented to achieve drift-free operation if the implementations of IDCT in the front-encoder and the end-decoder are known. In such a case, the decoder loop and the encoder loop can be implemented to produce exactly the same reconstructed pictures as those in the front-encoder and the end-decoder, respectively. Alternatively, if the implementations of the IDCTs in the front-encoder and end-decoder are not known, drift will not be significant, as long as the IDCT implementations satisfy IEEE Standard No. 1180-1990, which provides specifications for implementation of the IDCT, and macroblocks are refreshed as specified in other standards covering coding and communication of video signals, including the above-cited ISO/IEEE 13818-2 and ITU-T Rec. H.263. Since several coding parameters such as coding modes and motion vectors can be reused, the overall complexity of the architecture is not as high as the sum of a decoder and an encoder. However, the architecture is still computationally expensive.
To reduce the complexity of the cascaded pixel-domain transcoder architecture, several fast architectures have been proposed. For example, Eleftheriadis et al., “Constrained and general dynamic rate shaping of compressed digital video,” Proc. IEEE Int. Conf. Image Processing, Washington, D.C., October 1995, discloses an open-loop transcoder in which bit-rate adaptation is achieved by requantizing or truncating the DCT coefficients. Since the transcoding is carried out in the coded domain, where complete decoding and re-encoding is not required, the open loop transcoder can achieve fast operation. However, open-loop transcoding can produce significant quality degradation caused by drift due to mismatched reconstructed pictures between the front-encoder and the end-decoder.
To achieve less computation than the cascaded pixel-domain transcoder and better video quality than the open-loop transcoder, several other fast architectures have been proposed. Based on the assumptions that motion compensation (MC) and DCT/IDCT are linear operations, and a DCT will cancel out an IDCT (assuming the accuracy of DCT/IDCT operations is infinite), a fast transcoder architecture can be derived from the cascaded pixel-domain transcoder.
FIG. 2
illustrates the architecture of one such fast transcoder
200
disclosed in Keesman et al., “Transcoding of MPEG bitstream,” Signal Proc.: Image Commun., pp. 481-500, 1996. In
FIG. 2
, the functional acronyms within the blocks designate the same general functions as described above for FIG.
1
. The architecture in
FIG. 2
is based on the assumption that the same motion vectors and coding modes are present in the decoder and the encoder of the cascaded transcoder architecture so that they can be combined into one loop.
In
FIG. 2
, the DCT and IDCT are needed only for motion compensation performed in the pixel-domain. Therefore, if the motion compensation can be performed in the DCT domain, the architecture of a transcoder
300
illustrated in
FIG. 3
can be implemented as discussed in Assunçao et al., “A frequency-domain video transcoder for dynamic bit-rate reduction of MPEG-2 bit streams,” IEEE Trans. On Circuits Syst. Video Technol., Vol. 8, No. 8, pp. 953-967, 1998. This architecture is also disclosed in the above-cited Assunçao et al. reference and is referred to herein as the DCT domain transcoder. Thus, the DCT domain transcoder is configured by making substantial changes to cascaded pixel-domain transcoder architecture.
It would, however, be desirable to reduce the computational complexity of the cascaded pixel-domain transcoder architecture by means of less substantial changes in order to assure that the high picture quality this architecture provides is maintained.
SUMMARY OF THE INVENTION
Accordingly, the present invention is directed to a transcoder and methods of transcoding that substantially obviate one or more of the problems due to limitations and disadvantages of the related art.
Additional features and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be l
Lin Chia-Wen
Sun Ming-Ting
Wang Wen-Hao
Youn Jeongnam
Finnegan Henderson Farabow Garrett & Dunner L.L.P.
Industrial Technology Research Institute
LandOfFree
Transcoding apparatus and method does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Transcoding apparatus and method, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Transcoding apparatus and method will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3172818